The present disclosure relates generally to data analytics, and more specifically, to entity resolution between datasets.
Currently, entity resolution techniques compare data between traditional data sources, e.g., those in which the data is structured and includes strong identifying attributes. The resolution performs pairwise matching of data between the sources to determine candidate matches. If the datasets contain missing values, the process is often repeated.
The resulting data is disambiguated by a data steward who is employed to build entity groups from the pairwise comparisons. Decisions are then made on the cumulative evidence resulting from the disambiguation.